Impact of Detector Defects on Image Quality and Quantification for the microPET Focus 220 Scanner

Hardware problems with the detector systems of high resolution animal PET scanners could have an important impact on image quality and quantification. The aims of this work were to study the influence of detector defects on images acquired with the microPET Focus 220 animal scanner in terms of both, visible artifacts and quantitative accuracy, and to investigate possible approaches to ameliorate the effects of faulty detectors. Defects were simulated at the sinogram level by manipulating pixel values in the fully 3D sinograms that correspond to lines of response that involve the defective detector. Repair was also performed at the sinogram level using the constrained Fourier space method introduced by Karp et al. [1988]. Visible image artifacts were produced when a whole detector module (4 detectors) or a whole bucket (16 detectors) experienced total loss of efficiency. In addition, quantitative accuracy was decreased significantly in the case of a defective detector bucket with a bias in activity concentration of up to -36%. A defective module caused a bias of about -5% in most cases. A problem with a single detector had a negligible influence on the images and quantification. Sinogram repair worked effectively for a defective module. Very good quantitative accuracy was achieved for almost all data, and most artefacts in the images could be eliminated. However, in the case of a defective bucket, the gap in the sinogram data was too large for effective repair.